The ALICE Offline Alignment Framework



Offline alignment is the set of procedures combining all the sources of knowledge of the detector's geometry into the offline geometry, used in simulation and in reconstruction. While the hard-coded offline geometry is based on the detector's drawings, the real detectors'geometry can differ considerably because of limited mounting precision and time-driven deformations, such as twisting or sinking due to thermal and mechanical stresses. These deviations need to be corrected in order to minimise the loss of accuracy during conversion of signals into spatial positions and thus to increase the efficiency and precision of the reconstructed tracks and vertices, reducing as much as possible biased spatial information from the data. This is essential in particular for those physics analysis at ALICE which need to fully exploit the resolution capabilities of the central detector, such as the reconstruction of heavy quark particles'decays.

The corrections to the ideal geometry deriving from survey and alignment procedures are called alignment constants. Here and here we explain how the alignment constants are represented when passed to or saved into the alignment objects.

Here the functionality built into the framework is explained, which allows to store and retrieve the alignment objects in/from ROOT files, accessing them from the Offline Conditions Data Base (OCDB), and to apply them to the geometry. The explanations are accompained by some practical example covering the most common use cases. This functionality is particularly important in order to decouple the ideal geometry, produced by hard-coded function calls, from its corrections, which can than be stored separately in ROOT files and easily referenced.

The consistent application of the alignment objects to the ideal geometry held in memory requires to take into account some information which is out of the scope of the single alignment object; this is done automatically and transparently by the framework (see here).

The ALICE alignment framework is expected to be used:

  • to produce alignment objects;
    • from a macro, to simulate displacements;
    • from survey-to-alignment conversion procedures, which perform the conversion of survey data (positions of fiducial marks) into the corresponding alignment constants for the concerned volumes (alignment data);
    • from alignment procedures, which perform the optimisation of sensitive volumes positions based on specific physics signals, such as tracks produced during normal runs, lasers tracks or cosmic rays tracks.
  • to consistently apply alignment objects to the geometry in memory (in simulation and in reconstruction).

Simulation (digitisation) and reconstruction (tracking) need to query the aligned geometry to convert coordinates expressed in the global reference system into local coordinates and vice versa; these queries must be done according to the ROOT geometry package interface (usually referred to as TGeo). The ALICE alignment framework itself is based on TGeo ; this is taking advantage of TGeo functionality, such as storing and retrieving the whole geometry into and from a file or checking for overlaps and extrusions. In addition it gives the advantage of using the same geometry for simulation, tracking and visualisation, adding just a small "layer" to it for managing the additional alignment information.