DR=>IP<=DR Simulation of Linear Colliders with Ground Motion

Transcription

1 1of 37 DR=>IP<=DR Simulation of Linear Colliders with Ground Motion June 24, 2002 ISG8 Andrei Seryi, SLAC

2 2of 37 Needed a tool which would allow simulation of realistic behavior of LC and then learn using beam based alignment, tuning, etc. Tools and methods being developed by: K. Bane, L. Hendrickson, Y. Nosochkov, T. Raubenheimer, A. Seryi, G. Stupakov, P. Tenenbaum, A. Wolski, M. Woodley In this talk: Focus on simulations of dynamics effects like ground motion and feedbacks

3 BINP Measurements and development of hardware involve joint efforts of a lot more people - slow and fast motion studies Andrei Chupira, Alexander Erokhin Anatoly Medvedko, Mikhail Kondaurov, Vasili Parkhomchuk, Shavkat Singatulin, Evgeny Shubin BNL Nick Simos CERN - vibration propagation analysis - FD stabilization collaboration, etc. Ralph Assmann, Stefano Redaelli FNAL - slow and fast motion studies, girder stability, etc. Joe Lach, Chris Laughton, Duane Plant, Vladimir Shiltsev KEK - slow and fast motion studies, active tables, etc. Shigeru Takeda, Toshiaki Tauchi, et al. Northwestern University Mayda Velasco, Heidi Shellman, Inanc Birol, Gokhan Unel Oxford University - NUMI tunnel studies, stabilization - super fast feedback for IP Phil Burrows, Simon Jolly, Gerald Myatt, Gavin Nesom, Colin Perry, Glen White SLAC - slow and fast motion, cultural noises, stabilization, etc. Chris Adolphsen, Fred Asiri, Gordon Bowden, Marty Breidenbach, John Cogan, Domenico Dell Orco, Eric Doyle, Leif Eriksson, Joe Frisch, Linda Hendrickson, Tom Himel, Frederic Le Pimpec, Tom Markiewicz, Rainer Pitthan, Tor Raubenheimer, Robert Ruland, Andrei Seryi, Steve Smith, Peter Tenenbaum, Mike Woods, Nancy Yu 3of 37 Stanford University Sri Adiga University of British Columbia Tom Mattison, Russ Greenall, Parry Fung - turbulence induced vibration analysis - optical anchor development et al.

4 4of 37 Ground motion at diff. sites NLC site 127 shows very low ground motion (on surface!) As good as at LEP No cultural noises (yet) F, Hz Aurora mine is as quiet as SLAC site (may be impacted by inmine noises)

5 5of 37 Slow Ground motion Some data for diffusive or ATL motion: X 2 ~ A D TL (min.-month) (T elapsed time, L separation between two points) Place A µm 2 /(m.s) HERA R.Brinkmann,et al. FNAL surface * V.Shiltsev,et al. SLAC* Aurora mine* V.Shiltsev,et al. Sazare mine ~ 10-5 (1-10)*10-6 ~ 5*10-7 (2-20)*10-7 ~ 5*10-8 S.Takeda,et al. * Further measurements in Aurora mine, SLAC & FNAL are ongoing

6 6of 37 Ground motion models Based on data, build modeling P(ω,k) spectrum of ground motion which includes: Elastic waves Slow ATL motion Systematic motion Technical noises at specific locations, e.g. Integrated rms motion, nm "Model A" "Model B" "Model C" FD) 1E-4 1E Parameter A of the ATL, [ m**2/(m*s) ] in the models: (1, 5, 100) E-19 Frequency, Hz Example of integrated spectra of absolute (solid lines) and relative motion for 50m separation obtained from the models

7 7of 37 Mat-LIAR is our simulating tool Example of Mat-LIAR command files All features of LIAR, DIMAD, GUINEA-PIG + flexibility and power of MATLAB especially control tools needed for simulations of feedbacks Allow synchronous simulations of two beamlines pointing to each other their optics or beam parameters can be different

8 P(ω,k) is then used to generate x(t,s) and y(t,s) and beams GO IP 8of 37 Example of Mat-LIAR modeling

9 9of 37 P(ω,k) is then used to generate x(t,s) and y(t,s)

10 10 of 37 Important that correlation between e+ and e- beamlines is preserved IP Note that ground is continuous, but beams have separation at the IP

11 Simulations of complete NLC DR => IP <= DR 500GeV CM 1.98GeV 250GeV 250GeV 1.98GeV 11 of 37 Included: IP ground motion SLC style IP feedback RF structure misalignments Beam-beam effects

12 12 of 37 Intermediate ground motion

13 13 of 37 Zoom into beginning of e- linac Transition from tunnel filled with RF structures to transfer line

14 Zoom into beginning of e- linac 14 of 37 Transition from linac to transfer line

15 15 of 37 Noisy ground motion

16 16 of 37 Beam-beam collisions calculated by Guinea-Pig [Daniel Schulte]

17 17 of 37 Quiet ground motion

18 18 of 37 IP beam-beam feedback Colliding with offset e+ and e- beams deflect each other Deflection is measured by BPMs Feedback correct next pulses to zero deflection (it uses state space, Kalman filters, etc. to do it optimally) The previous page shows that feedback needs to keep nonzero offset to minimize deflection reason: asymmetry of incoming beams (RF structures misalignments=> wakes=> emittance growth)

19 19 of 37 calculated by Guinea-Pig [Daniel Schulte] Pulse #100, Z-Y

20 20 of 37 calculated by Guinea-Pig [Daniel Schulte] Pulse #100, Z-Y

21 21 of 37 calculated by Guinea-Pig [Daniel Schulte] Pulse #100, Z-X

22 22 of 37 calculated by Guinea-Pig [Daniel Schulte] Pulse #100, X-Y

23 With and without IP feedback, summary Example for one particular seed (seed is the same for the left and right plots) 23 of 37

24 More about IP feedback Design response curves for a step change for the IP feedbacks, standard SLC-style IP feedback (red dashed line) and the NLC design IP feedback (green solid line with circles). All are at 120Hz rep rate. Both SLC curves have 1 pulse actuator delay. There is no such delay in the JLC/NLC and CLIC designs (i.e. corrector will act on the next pulse). The blue curve with crosses corresponds to standard SLC linac feedback. More details can be found at the web page of Linda Hendrickson, at 24 of 37

25 NLC with ground motion B, several seeds. RF structures are randomly misaligned. Cases for both SLC-style and NLC design IP feedback are shown. Arrows show the mean luminosity for each seed. The arrows displaced to the right side correspond to the NLC design IP feedback. 25 of 37

26 Effects of fast and slow motion With only IP feedback, Lumi decays as orbit offsets at BDS magnets become too big. Have ~10000 pulses to fix orbits (should be quite enough) 26 of 37 Lumi decays as IP beam offsets become too big. The IP feedback fixes it. Simulations of slow effects are only possible with simplifications

27 Cultural noise at detector 1995 SLD measurements [Gordon Bowden] 30nm Measured ~30nm relative motion between South and North final triplets Magnetic field was OFF (magnetic field ON could have increases detector rigidity) North triplet (Ch1) noisier this side of the building is closer to ventilation and compressor stations Resonances (3.5Hz, 7Hz) are likely to be resonances of detector structure More quiet detector possible, but at what cost and how much more quiet? 27 of 37

28 Modeling detector vibration and FD stabilization 28 of 37 NLC with ground motion B, IP feedback and additional SLD-like detector noise (~20nm at each FD). Stabilization represented by an idealized transfer function.

29 29 of 37 Details of the modeling of FD stabilization

30 30 of 37 Ground motion IP feedback IP feedback, FD stabilization IP feedback FD noise IP feedback FD stab&noise A B 1.07 ±0.097 (12seeds) ±0.035 (4seeds) C Relative (to nominal) luminosity of JLC/NLC with ground motion, SLC style IP feedback (except several cases), FD vibration and FD stabilization. Averaged over first 256 pulses. Artificially imperfect machines were created by introducing σ x,y =(75,15) micron misalignment of RF structure with no further correction. Ground motion IP feedback IP feedback, FD stabilization IP feedback FD noise IP feedback FD stab&noise A B 10.4 ±2.3 (12seeds) ±1.2 (4seeds) C Pulse to the next pulse luminosity jitter (defined as RMS of (L i -L i+1 ) in percents to the nominal luminosity) of JLC/NLC with ground motion, SLC style IP feedback (except several cases), FD vibration and FD stabilization. Averaged over first 256 pulses. Artificially imperfect machines were created by introducing σ x,y =(75,15) micron misalignment of RF structure with no further correction.

31 31 of 37 For TRC we do similar studies with TESLA and CLIC

32 32 of 37 For TRC we do similar studies with TESLA and CLIC

33 33 of 37 Scan of one particular pulse of TESLA simulations, ground motion C, with additional noise at FD. The top picture show luminosity, the thin lines correspond to cases when either the vertical beam-beam deflection is zero, the IP offset is zero, or when luminosity is maximal. The bottom curve shows the vertical beam-beam deflection as a function of the vertical separation of the beams.

34 34 of 37 SR photons from beam-beam SR photons coming from the IP for ideal Gaussian TESLA beams (left) and for a one particular pulse of TESLA simulations, ground motion C, with additional noise at FD (left).

35 35 of 37 SR photons from beam-beam SR photons coming from the IP for some pulses of TESLA simulations, ground motion C, with additional noise at FD.

36 36 of 37 Simulations of complete NLC Tools developed. DR => IP <= DR We just started. A lot to test and to learn. For example, need to learn how to tune the machine with jittering luminosity And of course, simulations do not substitute developing hardware, taking more measurements, verifying models, etc.

37 37 of 37 For more details, see