## Session 3 — Operation & Modeling II | Title | Duration | Speaker | | --- | --- | --- | | Beam performance tracking & Beam Quality Monitoring | 20 | Amaury | | *Discussion* | 10 | | | Fault recoding, tracking & availability | 20 | Lukas (or Jean-Charles) | | *Discussion* | 10 | | | Improving tools and automation in the accelerator complex | 20 | Michael | | *Discussion* | 10 | | | **Coffee break** | 30 | | | Layout, optics, and operation: bi-directional integration | 20 | Kyriacos | | *Discussion* | 10 | | | Beam Simulations & Tracking: how good are our models? | 20 | Dora (or Elena de la Fuente) | | *Discussion* | 10 | | | Longitudinal Simulation tools: Where we stand? | 20 | tbd | | *Discussion* | 10 | | ### Beam performance tracking & Beam Quality Monitoring - Amaury - Developments throughout Run 3 - what we have, what we still need? - Do we need more instrumentation? How mature is it? (LBO, BGI, SPS BSRT ...) - Do we have control room tools suitable for day-to-day operation? - How to dectect degrated beam reliably and consistently and warn/alarm operators ... ideally on the onset - Evolution of BPT tools for run 4? ### Fault recording, tracking & availability - Lukas (or Jean-Charles) - Automatic Fault Recording developments across the complex in Run 3 - AFT in practice: quality and completeness of Run 3 fault entries, who fills them and when - Injector-chain coverage: is fault tracking as mature upstream as it is for the LHC? - What are the challenges for automatic fault recording and classification? ... could an AI / LLM agent help? ### Improving tools and automation in the accelerator complex - Michael - Inventory of what is already automated and the measured time saved - both classical and ML / AI - Settings management as the backbone - where can we improve - Where automation currently stops and why: gain vs cost, trust, interlocking, missing state information, ... - What must be in place before Run 4 restart (focus on infrastructure) - What would be nice to have for Run 4, what could be developed during the run? ### ~~Tools for hardware commissioning, LHC down the chain - Ola (or Cedric)~~ #### >>> more for recommissioning session? another talk on modelling? - Status of the HWC framework after Run 3: test coverage, automation level, and duration of a full campaign - Reusing LHC hardware-commissioning tooling for the injectors - Implementation of new test sequences and execution engines in AccTesting ### Layout, optics, and operation: bi-directional integration - Kyriacos (or Riccardo?) - Current chain from Layout DB → MAD-X sequence → LSA optics → operational settings, and where manual steps still break it - Consistency of naming, positions, and apertures across databases; how errors are caught today; what flexibility is needed, what is just an unnecessary source of errors? - MAD-X end of life: What will change with XSuite for run 4? - Can we have an online model (attempt #42 or so...?) - Optics Measurement and Correction strategies and tools - LHC and injectors - How to ensure layout information is updated "as-installed" and interface to installation - input from Giulia Romagnoli for input in Exp Area ### Beam Simulations & Tracking: how good are our models? - Dora (or Elena de la Fuente) - Benchmarks from Run 3 where simulation matched measurement well and where not - can better XSuite integration help? - tuning models to machines and correcting machines to models - online or near-online simulation for operational decisions: what is realistic today? - Validation strategy and MD requests needed to trust the models for Run 4 and HL-LHC - Input from Francesco ### Longitunal simulations and assessment (speaker tbd) - What models were put in place in Run 3, how well they described the effects? - Can we use the models for machine tuning? - What tools we have, and how do they integrate with transverse tracking tools like XSuite? - Check with Heiko and Helga