Choose timezone
Your profile timezone:
Optics repositories for current and future machines
→
Europe/Zurich
Optics repositories for current and future machines
Joint NDC-LNO meeting on Optics repositories for present and future machines, 09/06/2021
Introduction and news:
This meeting was organized by Rogelio and Stefano as joined effort with speakers from several ABP sections to discuss "Optics repositories for present and future machines".
Motivation : The number of databases, directories and repositories is increasing. Finding all relevant parameters to plan, study and operate machines can be difficult. The goal is to help to converge towards a single source that provides consistent sets containing all information required on optics including aperture, and to also provide for existing machines all what is needed to generate magnet settings and knobs for operation.
The last common NDC-LNO meeting has been on the 5 May. Minutes are on the indico page of this meeting https://indico.cern.ch/event/1033415/ . The connection cryostat has been realigned and will be re-checked after cooling. The inconsistencies in aperture information observed by Bjorn are being followed up and resolved.
In the meeting between directorate and experiments on Monday it was decided to limit the LHC magnet training to an equivalent of 6.8 GeV / beam with 100A margin. More on this meeting will be reported this afternoon in the LMC https://indico.cern.ch/event/1043312/
As the COVID sanitary situation improves, CERN is going slowly back to increase the presence at CERN for work in offices. Please keep EDH up to date accordingly and arrange to limit maximum occupancy of one person.
Status of the LHC optics repositories, Riccardo de Maria, slides
----------------------------------------------------------------
Historical and operational files have been on the shared file system AFS in
/afs/cern.ch/eng/lhc/optics
with sub-directories for optics versions 1 to 6.503 for the LHC design,
followed by RUNI, RUNII, .. directories for the running machine.
There are also directories for magnet errors and simulation tools, the LHC upgrades
SLHCV.., and HLLHCV1.0 to currently V1.5, HELHC and LHeC, the SPS and transfer lines.
The files are mirrored on EOS in
/eos/project/a/abpdata/lhc/optics
and linked from the web in
https://cern.ch/lhcoptics
More recently, git repositories are also used for the LHC. Riccardo says that git repositories store (efficiently) the complete history, that they can be hosted on a file system or on a website (gitlab.cern.ch, github.com), but would become rather unpractical for big sizes (> 100 MB).
Shared file systems can be used also for sharing large datasets. The AFS file system however is not supposed to grow anymore, and EOS introduces strong limitations and is not available on the technical betwork (TN).
https://gitlab.cern.ch/acc-models/acc-models-lhc contains operational optics files and optics models maintained by ABP and OP, and is fully operational for the generation of LSA (LHC control software architecture) settings.
A new web site is under construction to replace the current LHC Optics Web Home https://cern.ch/lhcoptics
Riccardo concludes that it is possible to migrate the complex LHC optics and related files entirely to the repository acc-models-lhc, but that this would require many changes to scripts and workflows and habits and could only be done slowly.
More work is needed to understand how to integrate magnet error information from FIDEL, WISE or JISE. We will hear from Piotr and Tobias on this front by July.
A common optics repository for the CERN (injector) accelerator complex, Alexander Huschauer, slides
-----------------------------------------------------------------------------------------------------
Alex reports that the new optics repository acc-models was introduced during LS2 for all CERN injector rings and lines, including AD, ELENA and CLEAR. It is accessible from the web as https://acc-models.web.cern.ch/acc-models/ and is based on the repository https://gitlab.cern.ch/acc-models. Previously separate BE-ABP and BE-OP optics directories were merged into the new repository. Alex shows the model team and illustrates the sharing of responsibilities, between our group, beam line, transfer line and operation teams. The new repository appears to already work well for all CERN machines (LHC not yet included) and be popular with users. The acc-models repository includes automatic checking, with running of MAD-X jobs and production of output and generation of plots. The repositories are also available on AFS as directory
/afs/cern.ch/eng/acc-models and EOS /eos/projects/a/acc-models/public/ . The EOS directory is at the same time the source for the website.
Larger data sets can also be linked directly from the web site. Information on SPS-aperture from several source is merged and will be included in the repository. Magnetic errors are not yet included.
Optics and Documentation Repository for FCC-ee, Ghislain Roy, slides
--------------------------------------------------------------------
Ghislain said the ideas were presented earlier in an FCC-ee optics meeting (12 March) and include the positive feedback he received in the meantime. Optics files, combined with documentation and links to papers and presentations are very important for FCC studies. Instead, temporary files and output files could stay on EOS but should not go to the repository. He illustrates the contents of the present https://gitlab.cern.ch/fcc-optics/FCC-ee-lattice and proposes a new naming convention like V21.02 where 21 stands for the year 2021 of creation. FCC-ee will not be ramped nor squeezed but should still allow for multiple files sets for each reference to allow for different energies, two beams, 2 or 4 IPs and thick and thin lens optics. The bv flag introduced for the LHC is not needed. While thin optics versions could in principle be generated "on the fly", the repository should also contain verified thin optics versions important as input for simulations for efficiency reasons.
The FCC is very large and merging all information in single files is not a good idea, and rather be separated into smaller files specializing on hardware, strength, or special parts like the solenoid. The repository will be complemented with sample jobs and automatic testing. Both MAD-X and SAD are used for FCC-ee optics. The testing should include automatic translation betweeen MAD-X, SAD (and back) to assure consistency. Tests should be provided by experts, like for example checks of aperture at collimators by the collimation team. Read access is provided to users based on egroups, and write access and git cloning only needed for developers. The repository should allow for extension to specialized studies on aperture, collimation, machine detector interface and impedance.
Riccardo comments that git based repositories allow for separate branches by individual developers that can be merged when generally useful into the master.
Leon comments that the insertion of correctors or the solenoid into SAD can be non-trivial and change the whole machine. He agrees that it would be good to allow for official thin lattice versions. Tapering should be done on the thick lattice before any slicing.
FCC-hh optics repositories and studies, Massimo Giovannozzi, slides
-------------------------------------------------------------------
FCC-ee and hh share the same main tunnel. The hh lattice was studied in detail for the CDR. It is stored in the repository https://gitlab.cern.ch/fcc-optics/FCC-hh-lattice
and also in /afs/cern.ch/eng/fcc/hh .
Following the outcome of the European Particle Physics Strategy group, more emphasis is now on ee as the first option to use the new tunnel.
Major changes to the FCC-ee like possibly going from 2 to 4 IPs that requires more symmetry in the layout will also require changes to the FCC-hh compared to the CDR version. Detailed placement studies for the FCC are currently in full swing with consequences for the injection lines and FCC-hh layouts. The corresponding FCC-hh optics with working files are currently located in /afs/cern.ch/eng/fcc/hh/PlacementStudies . There are also plans to shorten the dump sections from 2.8 to 2.1 km that require iterations on the FCC-hh optics. When the ongoing studies converge and the layout is consolidated, the files will move back to GitLab.
Optics repository for muons, Elias Metral and Daniel Schulte, slides
--------------------------------------------------------------------
Elias presented plans for the new CERN muon studies that launched few months ago. The study is organized in 9 working groups, with a leading role by Elias on beam dynamics. Information is available on https://muoncollider.web.cern.ch with links to latices, a readme.txt and literature on https://muoncollider.web.cern.ch/design/lattices . At this point, the lattice files are largely recuperated files from earlier US studies. Plans are to make all relevant files available as part of the general CERN Gitlab accelerator repositories discussed today and to extend this also to impedance information like https://gitlab.cern.ch/IRIS/LHC_IW_model . Muon accelerators do not just come as one ring, but rather as a combination of many layouts and complexes like drivers, capture and cooler.
Alex could provide help to integrate and link the Muon files.
Rogelio proposes to also include CLIC lattices with the help of Andrea Latina.
The timescale for the newer repositories containing FCC-ee and Muon files is to have first versions ready by July.
There are minutes attached to this event.
Show them.