Difference: AlfaOfflineDQMonitoringIntegrationToDo (29 vs. 30)

Revision 302015-07-16 - GrzegorzPawelGach

Line: 1 to 1
 
META TOPICPARENT name="ALFADataPrepToDo"

Offline DQ Monitoring Integration with ATLAS Central System

Line: 12 to 12
 

Manpower (people to contact)

Changed:
<
<
>
>
 

Goals

Line: 38 to 38
 
  1. Dead fibres (covered by 2D map of hits)
  2. Layer efficiencies (coverd by layer occupancy)
Changed:
<
<

2. List of distributions to be monitored

The following distributions shall be checked automatically or by the shifter for every run.

2.1. Number of hits per station

Purpose

Very basic distribution that can indicate problems although not necessarily point to their source.

Event selection

All triggered events used.

Description

8 histograms showing total number of hits per station. It is advised to remove bin representing 0 hits, as it may cause confusion.

Evaluation rules

The distribution is labelled with the following statuses/colours if the maximum of the distribution is in the range of hits:

  • good - [21, 24]
  • warning - [17, 21) or (24, 29]
  • error - outside of the range [17, 29]

2.2. Number of noisy fibers in each station

Purpose

Monitoring of hot fibers.

Event selection

Only events without trigger signal are selected.

Description

1 histogram with 8 bins. Each bin represents a station. For each station a number of fibres with signal in more than 1% of events without fired trigger is presented. This distribution is a summary of a 2D map of noise in fibres.

Evaluation rules

The distribution is labelled with the following statuses/colours if the number of noisy fibres for any of the stations is:

  • good - no noisy fibres in any of the stations
  • warning - between 1 and 5 noisy fibres in any of the stations
  • error - more than 5 noisy fibres in any of the stations

2.3. Number of inefficient fibers in each station

Purpose

Monitoring of inefficient/dead fibers.

Event selection

Only events with trigger signal are selected.

Description

1 histogram with 8 bins. Each bin represents a station. For each station a number of fibres with small efficiency, i.e. giving signal in less than 90% of events when it is expected to give a signal based on the reconstructed track, is presented. This distribution is a summary of a 2D map of fibres efficiency.

Evaluation rules

The distribution is labelled with the following statuses/colours if the number of inefficient fibres for any of the stations is:

  • good - less than 5 inefficient fibres in any of the stations
  • warning - between 5 and 20 inefficient fibres in any of the stations
  • error - more than 20 inefficient fibres in any of the stations

2.4. 1D distributions of occupancy per fibres layer

Purpose

Identify fibres layers with hot/dead fibers. Identify potential problems with PMT settings. This distributions provide overview of the detector efficiency. This distribution does not use tracks, thus is surely not biased by the reconstruction algorithm.

Event selection

All triggered events used.

Description

This is a set of 8 histograms, one histogram per detector. Each histogram is one dimensional and contains 20 bins, a bin per fibres layer. Both U and V layers are presented in the histogram. They are placed alternately and each bin is labelled with number and layer name. Histogram presents the efficiency of each layer which is calculated as a number of events with signal in at least one fibre in the layer divided by the total number of triggered events. This makes sure that the Z-axis range is from 0 to 1.

Evaluation rules

The distribution is classified according to the following rules:

  • good - efficiency of each layer is greater than 0.7 and difference between neighbouring bins is smaller than 0.2
  • warning - efficiency of each layer is greater than 0.6 and difference between neighbouring bins is smaller than 0.2
  • error - efficiency of any of the layers is smaller than 0.6 or difference between neighbouring bins is larger than 0.2

The values were selected by looking at distributions for run 267358.

3. List of distributions for experts

These distributions shall be prepared, but not checked for every run. They are to be used by experts to identify the problem.

3.1. 2D map of fibres efficiencies

Purpose

Identification of hot/dead fibres by eye-scan.

Event selection

Events with trigger signal and track reconstructed in main detector.

Description

A 2D histogram consisting of 64x10 bins is prepared for each U and V layer in each stations. This results in 16 histograms. Each bin represents fibre efficiency defined as number of hits in the fibre when the fibre is expected to give signal divided by the number of times the fibre is expected to give signal (eff = expected and observed / expected). The efficiency values are by construction is between 0 and 1. Fibres with less than 100 expected hits are masked with grey colour, because statistical uncertainty of their efficiency is large.

<!--
Evaluation rules

The distribution is classified according to the following rules:

  • good - no hot spots are spotted and tracks are clearly visible in all layers, there are no dead layers of fibers
  • warning - one or two hot fibres spotted and up to on dead layer
  • error - anything that looks strange or does not fulfil the good or warning conditions

-->

3.2. 2D map of noise in fibres

Purpose

Identification of hot fibres by eye-scan.

Event selection

Only events without trigger signal are selected.

Description

A 2D histogram consisting of 64x10 bins is prepared for each U and V layer in each stations. This results in 16 histograms. Contents of histograms is normalised to the total number of events, thus Z-axis range is between 0 and 1. A bar showing colour to number association is presented on the right side of each histogram. It is expected that most of the bins are empty.

Evaluation rules

The distribution is labelled with the following statuses/colours if there is no fibre with noise:

  • good - noise < 0.0035
  • warning - 0.0035 < noise < 0.01
  • error - noise > 0.01

The values were selected by looking at distributions for run 267358.

3.3. Distribution of number of hits per track

Purpose

Check if there is no strange behaviour induced by the reconstruction algorithm and if the layers efficiencies are high.

Event selection

Events with only one track.

Description

This is a set of 16 histograms (8 detectors x 2 layer types i.e. U and V). Each histogram is one dimensional and contains 10 bins. It is filled with the number of hits in a layer used to reconstruct a track.

<!--
Evaluation rules

The distributions shall be compared with reference ones.

-->

3.4. Number of tracks with more than 3 or 5 layers

Purpose

General performance of the detector, reconstruction algorithm and background monitoring. Allows to check e.g. ratio between events with trigger signal and without any reconstructed track.

Event selection

All triggered events used.

Description

16 distributions i.e. 8 stations x 2 tracks lengths (3 and 5 layers) of numbers of reconstructed tracks in an event.

>
>

2. List of distributions to be monitored and for experts

 
Added:
>
>
This list was moved to the official ALFA data quality site AlfaOfflineDQHistograms.
 

4. List of considered but not recommended distributions

Line: 282 to 50
 
Purpose
Changed:
<
<
Identify fibres layers with hot/dead fibers. Identify potential
>
>
Identify fibres layers with hot/dead fibres. Identify potential
 problems with PMT settings. This distributions provide overview of the detector efficiency.
Line: 301 to 69
  Second version: The set of 8 one dimensional histograms (8 detectors, layers U and V on one histogram), with 20 bins, a bin per layer.
Deleted:
<
<
Possible improvements

Shall this distribution be replaced with distribution of noisy layers or distribution of layers occupancy using all hits?

<!--/commentPlugin-->
 
Evaluation rules

The distribution of occupancy for layers used in track reconstruction

 
This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2019 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback