luke pickering prism weekly the aether 2019 01 10 l
play

Luke Pickering PRISM weekly The Aether 2019-01-10 L. Pickering 2 - PowerPoint PPT Presentation

Luke Pickering PRISM weekly The Aether 2019-01-10 L. Pickering 2 L. Pickering 3 Module wall/readout plane We have some off-axis events! Initial diagnostics threw up a few unknowns. Recorded POTs all wrong Non-Ar


  1. Luke Pickering PRISM weekly The Aether 2019-01-10

  2. L. Pickering 2

  3. L. Pickering 3 Module wall/readout plane We have some off-axis events! ● Initial diagnostics threw up a few ○ unknowns. Recorded POTs all wrong ■ ■ Non-Ar components in geometry need to be Cathode fidcuialized, leads to regions with no recorded POT ○ Mostly understood now: GENIE job seems to be failing, ■ Fiducializing Guang investigating. Some non-Ar features in Z, will ■ include in FV, but less important for POT weighting

  4. L. Pickering 4 Count FV POT exposure in ● 1cm slices: Use to ‘reweight’ to a ○ reference POT from any run-plan (or job failure…) Missing regions due to ● module walls, might be best not to move detector Exposure weight in exact meter stops. Non-Ar feature every ~0.5 m ○ Suggest 3.75 m stops. ○ Exposure weighting seems ● to work as expected!

  5. L. Pickering 5 Have some events! ●

  6. L. Pickering 6 Resurrecting old analysis code ● Getting there, still a number of steps to ○ make. Just true interactions in FV ■ Flux No reco selection/efficiency ■ correction yet. Fit weight code mechanically working, ○ seems to be some problems in the event rate prediction though. Incorporating some flux and xsec ○ variations possible if time allows once everything else works. CAF Event file Normalization doesn’t need special True numu CC ● & FV selection treatment anymore!

  7. L. Pickering 7 Cheating efficiency as before. ● Using 1.5 m vertex desert (0.5 for ○ standard FV + 1 m) Still see some structure on axis @ ○ high energy.

  8. L. Pickering 8 Ported Dan’s ● regularization optimizer Off-axis weights, Confusingly x-axis is x-pos * 2 to c++, and it ~seems to work. Choose reg strength To maximise bias/width trade-off

  9. L. Pickering 9

  10. L. Pickering 10 Far detector After numerous problems with the ● beamsim, things seem to be working now. docdb:1517 Thrown almost all of the variations that had ● any effect in previous studies. Results look very similar to previous ● studies, which is good, as not too much has changed: Three horns, Lower power proton driver ○ One missing feature, still investigating. ○ Hoping to present at BIWG before CM, but ● Laura seems mostly happy with status.

  11. L. Pickering 11 docdb:1517 Main obvious missing feature can be ● seen in the WSB in both beam modes. Far detector I have run a few more tweaks (include ● HC to - 1sigma) and will include these this afternoon. Hopefully that is the last piece and I ● can test and send out these flux uncertainties in the next few work days. Test in CAFAna and in linear combination fits. ○

  12. L. Pickering 12 Many more details here. ● Expecting plots like this to make it into ● the TDR.

  13. L. Pickering 13

  14. L. Pickering 14 Final flux uncertainties: ● Need to present to BIWG, hoping to do that early next week. ○ Produce 6 fits for Fake Data study ● (With, Without off-axis) x (CVN, Parametric, Parameteric + fake data) FD Recon ○ Me/Guang to make sure it all works ■ Callum will do final, ‘official’ fits ■ Used to inform ‘systematic’ to cover possible bias in dcp, should dramatically ○ reduce sensitivity. Argue DUNE-PRISM leaves us less susceptible to such problems. Produce linear combination prediction from CAFs, with/without fake ● data shifts. Currently trying to get this done with standalone code ○ I don’t anticipate having time to work on full CAFAna disappearance analysis, Will ○ Guang, or are we punting for now?

  15. L. Pickering Thanks for listening

Recommend


More recommend