Availability, Operation & Energy (some preliminary thoughts)

1. Availability, Operation & Energy(some preliminary thoughts) Paul Collier, CERN FCC I&O Energy

2. Some Preliminary Ideas for discussion on: • Availability : • Maximizing the time the machine(s) are able to take beam • Minimizing down time • Operation : • How the machine(s) will operate – over the years • An operational Cycle … how we will do Physics • Energy : • What kind of electrical load is it likely to represent • Identification of major energy consumers for optimization Most of the talk will concentrate on the hadron machine (FCC-hh) some additional comments concerning the specificities of FCC-ee Discussion based on recent LHC experience FCC I&O Energy

3. Operational Cycle Beam dump Squeeze Stable beams Energy Collide Setup (Ramp down and Preparation for next fill) Ramp Injection Should take this as a basis for FCC-hh as well and see the impact FCC I&O Energy

4. FCC-hh Impact of Operational Cycle: 2h to fill, 10h fill length • Injection • LHC can be filled theoretically in 8 min. • In practice with present injector cycles is takes 30 mins (minimum) • But for FCC, 4 times the number of bunches and a new injector! • Transferring large beam energies between machines!! • Filling FCC in 10-30 mins has a big impact on the injector design • Ramp • FCC-hh will have a similar energy swing as LHC (~ x16) • Maximum ramp rate in LHC 10 A/s • Maintaining a similar ramp time in FCC-hh has big implications on: • Powering Sector length – giving total inductance in the main circuits • Magnet design if higher voltages are needed and for magnet inductance • Both of these have implications on stored energy in the magnets and powering protection systems • Similar considerations for ramp down and squeeze FCC I&O Energy

5. FCC-ee – Operational Cycle The short beam lifetime in collision requires “top-up” operation Collider ring runs in DC at the physics energy The Accelerator ring cycles at ~0.1Hz • Theoretically the efficiency can be 100% • The last stage in the injector is included as part of the design. • However FCC-ee will be much more to exposed to faults in the Injector complex • LHC (FCC-hh) Can ride through some injector downtime by staying longer in physics. FCC I&O Energy

6. FCC-ee Power Consumption Estimates Based on the 80km Machine Study – Should not be too different to a 100km version. Pre-injectors not included. It includes the infrastructure scaled to the need for TLEP and not that which would be installed to allow a future installation of a pp machine. I wont say more about this today .. The Key Driver here is the RF system: Cavity characteristics and efficiency of the RF power sources (assumed 55%) FCC I&O Energy

7. FCC-hh Power Consumption To first approximation: Most will scale to FCC-hh very approximately according to length (ie x4) The Experiments are likely to be more than LHC but not by a large factor • Beware: This is a ball-park figure to set a rough scale!! • Clearly the Cryogenics is a key driver • But the infrastructure itself (cooling/ventilation) will also be a large consumer • The RF system itself (if >60MV is needed) is significant • same R&D for ee and hh machines !! FCC I&O Energy

8. How to Refine the FCC-hh Estimates In the absence of some detailed design of the systems the best bet is to try and refine our estimates of scaling from LHC The initial estimate was based purely on the length scale – ie 4xLHC For each major system we need to refine this to look how the scaling should be done Or, better still, make some preliminary estimates based on the actual FCC. Some preliminary thoughts here mainly on the magnet powering … will need more input FCC I&O Energy

9. FCC-hh Magnet Powering Main Magnet Sectorization • LHC Has 8 arcs of 2.9 km. Each one a powering sector in each there are 3 main circuits – 1MB and 2MQ. • Assuming FCC-hh has 12, 6.9km length arcs we can make 2 assumptions: • One powering sector per arc : i.e. 2.4x the length in the LHC • Two powering sectors per arc : i.e. 1.2x the length in the LHC • In each case assume that the magnets have the same characteristics as the LHC • The voltage required for each circuit can be split into 2 parts: • Steady state voltage to overcome the resistance in the circuit • To first approximation this does not depend on the length of the circuit • Peak voltage required to change the field – up or down. • To maintain the 10A/s ramp rate of the LHC this voltage must increase as the inductance in the circuit increases • Alternatively the ramp-rate will decrease • Assume that we want to keep the present LHC ramp rate for the FCC-hh. FCC I&O Energy

10. The FCC/b scenario is probably the most likely … because of stored energy considerations. In addition, the time constant of the dipole chain in the a) option will be over 50,000 seconds (LHC = 23,000s) FCC I&O Energy

11. Power Converters Steady State Power .. • Scaling for Steady power = 2.9 based on 12 arcs – rises to x3.1 if we split them (above). • Assumptions made for the LSS have a significant effect • h-h option presently assumes 4x LHC in total length (I assume same cell length) • I Keep 2 LSS for beam cleaning – but twice as long • The dump septa is scaled by the energy and becomes enormous 8.4MW! • Clearly a technical solution using SC septa would be lower. • Note : PC power output considered here, not mains power! • Total of around 6300 power converters and 63MW! FCC I&O Energy

12. Cryogenics (L. Tavian) Per arc For FCC-hh (12 arcs) State-of-the-art cryoplant LHC installed power LHC cryoplant w/o cryo-distribution ! w/o operation overhead ! L. Tavian - 22 May 2014

13. OtherMachine Equipment Cryogenics Estimates made at the kick-off meeting – 140-190 MW power to refrigerators depending on temperature of the cold mass. Assume 1.9K. Gives a factor >4 compared to LHC Cooling Presumably a function of the total power to remove – will therefore scale with the scaling of the total consumption? Ventilation I assumed only length – presumably it should more properly scale with volume. Since the tunnel section will be larger it might even be greater than x4. RF Initial estimate 4xLHC (determined by beam stability). More refined estimate points to ~2.5-3 However Lower/Higher harmonic cavities & crab cavities would increase the power needs For the moment ignore them – but bear it in mind. Input from Machine needed. Other Machine A factor 4 looks reasonable – but some things might scale with Energy!! FCC I&O Energy

14. The Rest • Experiments • Seems reasonable to assume they will consume more – but not that much • Experiments are dominated by electronics, I think • First ideas from the experimental team would help a lot here. • At the moment I assume ~1/3 more power • Injectors and Transfer Lines • Clearly nothing done here yet • Likely that (part of) the transfer lines will be powered from the FCC infrastructure • General Services • Were left off the preliminary estimate! • Complicated … scale with number and size of buildings, number of shafts, total length – depending on what is being powered. • Stick with x4 for the moment • General Services in LHC = ~20MW • Therefore FCC would need ~80MW !! FCC I&O Energy

15. New Estimate – It is not getting better! Anything else I have forgotten? At present PC’s are output power – need to estimate mains power. Do we need to consider transmission losses for the power distribution? FCC I&O Energy