Flowdown notes January 15, 2019

From CMB-S4 wiki
Jump to navigationJump to search

JC -- overview of goals for this call and relationship to the Friday DSR call and the bi-weekly Science Council call.

We received very helpful feedback from the DSR review in December. The need to better develop the science to measurement to instrument flow down needs to be much better developed. For the DSR reference design we do not need the level of traceability that we will eventually need for CD-1, PDR, etc, but we certainly need to justify the science to measurements requirements and a good, credible start to the measurement to instrument flow down.

The plan for the DSR is to put this flowdown details/justification into an appendix, and then have a short chapter (Chapter 2) in the main text giving the summary. We will further develop the appendix -- increasing the rigor -- as necessary for upcoming reviews after the DS.

Note that the expected date for initial project papers from the Decadal Survey was announce to be May or possibly June at the AAS Decadal Survey Town Hall. That said we want to have the full draft version of the DSR flowdown ready to review at the March 13-15 CMB-S4 workshop to be held at FNAL.

On other telecons: The Friday DSR telecon is focused on the project management. How do we build the reference design? Cost, budget, system engineering, etc. The science council telecons (every other week, the off-weeks of this telecon) are focused on the science case and its flow down to measurements.

Lloyd: The SC is working to advance the science case for a compelling DS A&A case. That is chapter 1 of the DSR. They will then come up with a ~6 level 1 science goals that drive the measurements, more than just r and Neff as was done for the CDT. They have asked the individual SC working groups to each come up with ~3 such goals for next week.

Natalie -- see presentation - https://docs.google.com/presentation/d/1BmzEnZ7NH0blrFJhaXI8rSH44GB-PH06oJgBDbrB3CM/edit#slide=id.g4d7c0b3f61_0_45

Main requirements for DSR

  • L1 Scientific Requirements define high level scientific objectives
    • drive the L2 Survey Requirements, connected by simulations and judgement
  • L2 Survey Requirements define the survey to be carried out
    • drive the L3 Technical Requirements, connected by optimization studies, simulations and judgement
  • L3 Technical Requirements define high level technical choices
    • Can include constraints, eg assumed number of operating years, weather, etc
    • Flow down to detailed L4, L5 requirements on the instrument & software

Note that this is an iterative process. Might saturate a science goal, might saturate cost, etc. and require iteration.

It is living document from design phase through commissioning. Used for defining instrument to assessing as-built performance to science goals.

Walked through level 1, 2 and 3 using CDT as an example.

Requirements Chapter for DSR:

  • Use existing tools to describe connections between Level 1, 2 and 3
    • r forecasting codes
    • Optimization tools
  • Allow for some iteration in goals
    • eg is Neff goal achievable?
    • Which r goal is actually driving the design?
  • Make “reasonable” assumptions (eg foreground model, de-lensing capability)
  • Must be able to extract requirements and assumptions (eg efficiency factors, scan strategies, noise, resolution, etc) from the simulations and include in req document

Discussion: Currently we have lumped various yields into "achieved" results, but we should understand the individual pieces and compare to what has been achieved. We have to be careful not to fool ourselves. There is death by a thousand cuts, and we would not be well served by only considering a few of the biggest ones. Need to have strong cross-check back to achieved performance.

This will require substantial work, especially if we want to be able to compare impact of the sites.

We need to show optimization to achieve goals, but also cost optimization.

Nadine: System engineering, starting at slide 15. Near term is to flow to our main sub-systems.

Zeesh: L1, L2, L3 spreadsheets being populated so far by the CDT and DSR.

ACTION ITEMS: Zeesh and Nadine to start next steps on slides 21 & 22; John Kovac to start on basic efficiency factors;