-- HelenHayward - 16 Aug 2021

Introduction

This page is for the collection of ideas, draft proposals and list of current and past projects in collaboration with the ATLAS Upgrade Group

Project Ideas

Please add here ideas for projects for students. Although primarily for MSc and BSc students, we should also consider Qualification tasks. (note: 230821: After speaking to Monica, it appears we can not offer projects to LivDat students, as they must be done in industry).

  1. Improved Environmental monitoring of LSDC (Sven)
    1. create an Influx DB and Grafana monitoring on the Linux cluster
    2. commission REMS temperature/humidity monitoring with data output to Influx DB
    3. comission particle counter with data output to Influx DB
    4. survey of cleanroom to identify places where temperature/humidity monitors and particle counters should be added
    5. add aditional monitoring devices (if they are available)
  2. Mechanical Analysis (D Sim?)
  3. ML for visual QC (Paul)
    1. splitting out hybrids from panel automatically
    2. cutting the strip hybrids in half automatically
    3. automating the whole process from a scanned image --> lots of steps (done by hand for now) --> result and database upload
    4. to make it a bit more interesting they could investigate other object detection methods (Mask-RCNN etc....) to compare to YOLOv5
    5. generating "real" looking fake images using a GAN network would still be very nice to have (they use this for x-ray images of covid affected lungs, so its used in other areas).
    6. Then there is the planning for pre-production/production. So turning an idea (yolo and my cnn) into a production system. Thats updated as it runs (retrain the networks once a week/month). It's called MLOP's (machine learning operations). This bit is more long term project and might suit a LivDat person.
  4. mechanical analysis-
    1. manual measuring of machined componants=carbon foam & machined radius using existing jig, dowels and height gauge.
    2. stave deflection data comparrison, trends etc.
    3. stave componants mass correlation,
All the above will be relavant as a whole batch of componants will be here for testing in preparation for the first PPA stave such as c channels, facesheets, lock points, honeycomb, cooling loops and freshly machined foam. All componants can be quality controlled with existing techniques with plenty of previous data available for comparrison.
Topic revision: r6 - 08 Sep 2021, SvenWonsak
This site is powered by FoswikiCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Foswiki? Send feedback