Minutes from 29 Mar 2006 SWMD Telephone Meeting

Notes from previous meeting

1. Material damage
   • Conclusion of SLAC damage test work - Justin (for George)
     George had studied temperature as function of time over short period 0.01s to check behaviour using smaller time steps - noted that ANSYS calculations are cumulative, so uncertainties at very short time scale will lead to unreliable results at longer timescales, even if the behaviour looks like aymptotic convergence to some physical value. Used energy deposition in space from one of Luis' Fluka simulations at 250 GeV, goal is to determine appropriate time steps to reduce uncertainties to acceptable level. Large fluctuations found for time steps 2e-4s, 1.e-4s, smooth reduction in temperature when steps of 25microsec used. Due to boundary conditions imposed, the max. temperature rise is artificially less than expected at long times - heat cannot escape from the surface of the model. In ESA, only a relatively small amount of energy could be deposited in the collimator material in a very small volume, after this has spread even a short distance, the max. temperature rise would be dramatically reduced. Even with boundary condition limitations, peak temp. rise of ~250K is reduced to 50K after 5ms - so conclusion is that 10Hz beam would not offer a realistic test.
   • von Mises stress for the fully constrained model - Justin (for George)
     Simulation of VM stresses with time, showed clearly stress wave reflected from surface of volume, not so obvious why stress is non-zero at small initial times, as definition of VM stress requires some asymmetry in components resolved along three orthogonal axes, and initially would expect equal expansion in all directions. Can be checked by examination of individual stresses along each axis. At end of simulation, would be interesting to see behhaviour at slightly longer time when stress wave re-enters the initially-heated zone, e.g. additional 8e-12s?

2. Wakefield simulations:
   GdfidL output for cylindrically symmetric jaws
   1. Summary of pbx runs
   2. Summary of pbxb runs
   3. Runtime durations
   
   Jonny described the detailed studies he has carried out on sensitivity of his loss factor simulations to input parameters such as no. of mesh cells per sigma_z, aspect ratio of the mesh, and bunch length, using as example the pill box cavity. Discussed for transverse loss calculation what the appropriate normalisation is for the case when GdfidL chooses not to carry out integration along the requested displacement off the symmety axis - does transverse loss factor (V/pC/m) use distance *requested* off axis of symmetry, or *actual* distance? Appears that results are for "requested". Results are found more stable for larger cells/sigma z, results are stable to within 10% relative uncertainty using 8 cells/zigma. For mesh aspect ratio, choice will depend on model being studied to some extent. Aim is to include ABCI calculation on his parameter sweep comparisons (already includes analytic calculation) and then continue with collimators for ESA run - using default 8 cells/sigma z - and single choice of aspect ratio, and 0.3mm sigma_z. Additional parameter variations to demonstrate stability around these values can be studied after baseline predictions have been made. Noted that collimators 7 and 8 will be different to others having two different taper angles and collimator 3 will have few different flat section lengths studied, not expected to be necessary to include 1m section.

3. Analytic calculations - spreadsheet - Nigel
   Minor changes to this to include predictions of deflection angles per unit offset of beam/collimator centre (microrad/mm). Deflections expected to be substantially smaller than in previous runs, so need to average over more bunches to achieve good resolution on kick factors. Needs stability of BPM calibration over at least the time of a signicant part of the scan. Implication for scan strategy (e.g. high statistics points throughout collimator scan, perhaps recalibration, then filling in remaining points?).

4. Wire/bench test preparations
   Carl had noted yesterday that there was a delay in hardware fabrication required for the tests, 3-4 weeks.

5. Progress for ESA tests
   1. Arrived safely at SLAC - 21-Mar-2006 - now with vacuum group.
   2. Sandwich modifications for the new Uk collimators are being made, Steve Molloy is taking charge of these
   3. Matlab Reconstruction code - Nigel (for Luis)
      Luis has made a first simulation of the effects of finite BPM resolution on the measurements, taking BPM z locations from optics tables (two upstream BPMs, c/o Frank J.), and of six downstream BPMs from the layout plans (therefore approx. - not from a surveyed measurement). He has assumed 1 micron rms noise on each BPM, and with 50 pulses/y position the resolution on kick factor is less good than expected, for 500/position looks OK, with relative error on kick factor (of 0.5microrad/mm) of ~10%. For this, a purely linear dependence on the kick angle as function of y was assumed for simplicity, this can be made more sophisticated (e.g. cubic near collimator wall) at later stage.
      Next steps are to include beam jitter, implemented as coherent noise assigned to all BPMs.

   4. Travellers for ESA run
      Please confirm travel plans with Mike Woods, cc: Nigel if you have not done so already. Also, could people travelling to SLAC for the run, please try to phone in to the ESA Monday meetings for the next few weeks if they do not do so already.

6. Future plans
   There will be a (closed, management/WP/task leaders) UK meeting on 3-Apr to discuss aims for post-Apr-2007 period, with open meeting at Manchester on 24-Apr.

7. Next meeting: date to be fixed outside meeting, suggest Thu. 20-Apr