the 2011 tracker misalignment scenario

15
R. Castello (UC Louvain) on behalf of the Tracker alignment group The 2011 Tracker misalignment scenario CMS AlCa meeting Cern, 22/11/2011

Upload: lane-pope

Post on 02-Jan-2016

39 views

Category:

Documents


0 download

DESCRIPTION

CMS AlCa meeting Cern, 22/11/2011. The 2011 Tracker misalignment scenario. Motivations for a new scenario. Previous scenario (TrackerAlignment_2010Realistic) provided a worse description of the real remaining misalignment and global deformations on data - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: The 2011 Tracker misalignment scenario

R. Castello (UC Louvain)

on behalf of the Tracker alignment group

The 2011 Tracker misalignment scenario

CMS AlCa meetingCern, 22/11/2011

Page 2: The 2011 Tracker misalignment scenario

Motivations for a new scenario

Previous scenario (TrackerAlignment_2010Realistic) provided a worse description of the real remaining misalignment and global deformations on data

Aim: a new one which performs closer to the current data alignment in view of 2012

Mostly in: DMR/residual distribution in the Endcap momentum biases in resonances decays (mass VS muon phi)

R.Castello 2

Page 3: The 2011 Tracker misalignment scenario

Input dataset and studies performed

Current alignment on data performed mainly with MinBias+Cosmics A balanced input of ALCARECO (MC):

Isolated muons from W (16.3 M tracks) Isolated muons from QCD-mu enriched events (2.3 M tracks) Minbias tracks (450k events ~ 4.5 M tracks) Zmumu events using mass constraint (390k events) Peak mode cosmics (1.8 M used tracks) Deco mode cosmics (1.7 M used tracks)

Opportunely re-weighted to match the pT//phi track distribution on data

Several strategies studied: Starting from Startup MC + bow parameters from data Starting from ideal MC + small random noise + no bow mis-alignment Starting from ideal MC + bow parameters from data different weigh to the cosmic sample

Lesson learned: Including bow misalignment hit CPU limit: asked for an increasing The remaining Z mass bias versus muon phi seems to be completely

dominated by starting geometry (startup has a too large one w.r.t. data): intermediate starting point

Weighting cosmics (x2, x0.5) has almost no impac: discareded for now

R.Castello 3

Page 4: The 2011 Tracker misalignment scenario

Final strategy

Final solution was to create an “intermediate” starting geometry Module by module differences between Ideal and Startup scenario

extracted (ID-ST) Differences applied on top of Ideal MC sceanrio and used as starting

point for alignment procedure with the input dataset described Three sizes:

Ideal02= ID-ST x 0.2

Ideal03= ID-ST x 0.33

Ideal05= ID-ST x 0.5

Bowed sensors param from data included as misalignment

on top of the starting geometry

Same alignment strategy usedfor data (one iteration)

R.Castello 4

ID vs Startup ID vs Ideal03

Page 5: The 2011 Tracker misalignment scenario

Validation

R.Castello 5

Page 6: The 2011 Tracker misalignment scenario

Bowing deformations

R.Castello 6

Page 7: The 2011 Tracker misalignment scenario

DMR (Barrel)

Previous scenario already performing good in the Barrel: small changes

R.Castello 7

Ideal MCdata geometryStartup MC scenarioNew MC scenario

Page 8: The 2011 Tracker misalignment scenario

DMR (Endcaps)

New MC scenario better matches the data in the Endcap Last improvements in the data alignment are better described by

this scenario

R.Castello 8

Ideal MCdata geometryStartup MC scenarioNew MC scenario

Page 9: The 2011 Tracker misalignment scenario

Mass Vs muon phi

Remaining amplitude (05Delta) closer to data now, w.r.t. previous MC Startup scenario

R.Castello 9

Page 10: The 2011 Tracker misalignment scenario

Mass Vs muon phi (Barrel/Endcap)

Different behavior between Barrel EndCap Closer performance to data in the Endcap, Barrel still a bit optimistic.. Starting geometry “subdetector dependent” has been produced for

accounting this, but no time to validate and sign-off

Potential candidate: MC_idealplus05Delta

R.Castello 10

TEC -Barrel

Page 11: The 2011 Tracker misalignment scenario

Z mass peak

Good matching in the Z mass shape and resolution

R.Castello 11

McS

cean

rio_Ide

al05

IDEAL vs Mcsceanrio_ID05 Data vs Mcsceanrio_ID05

Caveat: not normalized

Page 12: The 2011 Tracker misalignment scenario

Validation in other variables: mass vs (_mu1; _mu2)

R.Castello 12

Overall shift, but trend more similar to data

Page 13: The 2011 Tracker misalignment scenario

Cosmic track splitting

Not noticeable differences visible new scenario (BPIX cent) remarkably similar to IDEAL

R.Castello 13

Ideal MCStartup MC scenarioNew MC scenario (BPIX centered)

Page 14: The 2011 Tracker misalignment scenario

Summary

A new MC misalignment scenario has been produced for the new MC production: Overall the behavior of the new scenario is quite good in describing the

data Close to the ideal geometry! It includes bowing deformation fitted in dataReady to be used: already uploaded in the db (PXB centered), seehttps://hypernews.cern.ch/HyperNews/CMS/get/calibrations/983.html

Still room for improvement: Study on-going for a MC scenario that better matches the data

performances in the Barrel region BUT AlCa+Offline groups had the urgency to move on with the MC

production asap to meet the deadlines imposed by the Physics groups.Payloads:TrackerAlignment_2011Realistic_v1_mcTrackerSurfaceDeformations_2011Realistic_v1_mc

Credits for this work : A.Agocs, J.Behr, A.Bonato, R.C., G.Flucke, N.Heracleous, K. Krajczár, G.

Vesztergombi R.Castello 14

Page 15: The 2011 Tracker misalignment scenario

BACKUP slides

R.Castello 15