Scheduling the James Webb Space Telescope at L2

1
Scheduling the James Webb Space Telescope at L2

• Knox S. Long,
• John C. Isaacs, Wayne Kinzel, Larry Petro, Peg
  Stanley Hervey S. Stockman
• SpaceOps2006
• Rome

2
The Observatory

• IR-optimized telescope, passively cooled to 37K
  intended to operate for 5-10 years.
• Segmented primary that must be actively
  maintained
• 4 science instruments as complex as HST
• Autonomous communications to DSN SOC
• 1 contact per day, 200 GB per day compressed
  (21)
• 2 days on-board storage
• NASA/ESA/CSA

3
Location at L2

• Allows passive cooling
• Large solar shield blocks light from sun, earth
  and moon
• Simplifies operations
• Long uninterrupted viewing
• Few mandatory interruptions in observing schedule
• Station-keeping every 22 days

4
The Science Instruments

• NIRCam (Univ Ariz)
• 0.6-5 µm imaging
• HgCdTe detectors (40 Mpix)
• WFSC
• NIRSpec (ESA)
• 0.6-5 µm Multi-object, long slit, and IFU
  spectroscopy
• HgCdTe detectors (8 Mpix)
• MIRI (ESA/NASA)
• 5-28 µm imaging
• Slit and IFU spectroscopy
• SiAr detectors (2 Mpix)
• FGS-Tunable Filter (CSA)
• (R100) tunable filter imaging
• HgCdTe detectors (16 Mpix)

5
JWST Science Themes

• End of the dark ages first light and
  reionization
• The assembly of galaxies

The Eagle Nebula as seen by HST

• Birth of stars and protoplanetary systems
• Planetary systems and the origins of life

Galaxies in the UDF
6
A General User Facility

• HST after 15 years
• 4138 unique PIs and Co-Is
• 3913 selected proposals
• 5781 unique authors of refereed papers
• 7300 registered users of HST data archive.
• Webb will be similar
• Expect 1000 proposals per cycle
• 100-200 selected proposals per cycle, after first
  year with a TBD share of Legacy proposals
• gt85 of time awarded to general observers

7
Challenge Best science within constraints!

• Main architectural constraints
• Field of regard
• 40 of sky in instataneous FOR
• Pitch 5 45 deg,
• Roll ? 5 deg
• Solar radiation pressure loads reaction wheel
  assemblies
• Momentum dumps must be limited
• Use fuel
• Make orbit tracking more difficult at L2
• Station-keeping (SK) every 22 days

8
Science constraints

• Science policy constraints
• A small (1.1) oversubscription of available time
• Time variable backgrounds ? ltS/Ngt within 5 of
  best
• Commitment to complete approved program (90
  within a year)
• Science program characteristics
• Location and exposure time distribution of
  sources
• Timing and orient constraints of observations
• NIRSpec observations (may) require prior NIRCam
  snapshot

9
Webb Operations Concept
JWST

• Users submit flight-ready Phase II proposals
• GO, GTOs, and staff use same proposal tool
• STScI generates a long range plan (LRP) for year
• Time periods for scheduling
• STScI generate short term schedule from LRP

• STScI generates OP segments from schedule and
  sends to Observatory
• JWST executes observations and returns data
• STScI updates LRP to reflect success or failure

10
Long Range Planning

• Sets time windows for all approved science,
  calibration and maintenance visits to allow
  science requirements be met
• Balances critical resources, such as fraction of
  hard visits
• Creates a time scale for completing all detailed
  planning
• LRP process remove visits as they execute, adds
  new visits as approved, accommodates changes in
  existing visits

11
Short-term scheduling

• Consists of establishing the exact order of the
  observations
• Does not establish the absolute time for start
  and stop of most observations
• Webb observations will, instead, be event-driven
• Unlike HST, Webb will not wait until the next
  observation on a guide star failure
• Instead it will move to next observation

• Event-driven means that there is a queue of
  activities that are executed sequentially
• An activity is started and returns a message at
  successful (unsuccessful) completion
• When the first activity ends, the next activity
  begins.
• Exact times to begin and end are not pre-ordained
• The only option the S/W has is to start or skip
  the next activity in a list of activities

12
Successful Plan Execution
Visit 1
Visit 2
Visit 3
13
Failed Guide Star Execution
L latest start time
14
Simulations of Scheduling

• SO-DRM JMS
• Specific set of programs that reflect the high
  priority science programs of JWST
• Software that can accept these programs in their
  full complexity and schedule them
• Represents natural and observatory source of
  exposure noise ? S/N studies possible

• Monte Carlo (planJ)
• Monte Carlo simulation large variety of potential
  science programs specified with less fidelity
• Distribution of targets and exposure time
• Fixed orient only
• Software that more fully implements planning
  process or long and short-term scheduling

15
SO-DRM from Science Req.

• 3,715 primary visits
• Visits have short duration, i.e., ltlt target
  visibility
• Average visit duration is 0.13 days
• Average time at given place on sky is 0.31days
• Most visits (91) are constrained
• 33 have orientation constraints
• 11 have relative timing links
• 42 have groupwithin links

16
Simulations show high utilization with small
oversubscription

• SO-DRM program can be scheduled with high (gt95)
  1.1 oversubscription
• Scheduling to maximize S/N possible
• Monte Carlo tools similar results for a wide
  variety of programs

17
Momentum Management Challenge

• Longest observations with JWST are expected to be
  of order 10 days
• At some orientations, solar torques load RWA in a
  few days
• Design of sunshield and RW capacity has changed
  since 2005 Sept
• But ratio lttgt/M about the same
• No more than 2 momentum dumps within Station
  keeping intervals
• RW loading must be managed via scheduling

18
Parameterizing momentum management
Expected Momentum L sqrt(n) lttvisgt lttgt

• Models torque build-up as 2-d random walk
• Dump-time (trw) determined by
• momentum capacity (M)
• visit time (tvis)
• torque buildup rate (t)
• Architecture factor P indicates impact of
  momentum management on scheduling

Predicted dump time lttrwgt M2 /
(lttvisgtlttgt2) where lttgt2 0.5
(ltt?1gt2ltt?4gt2)
Architecture factor P
sqrt(lttdumpgtlttvis) lttgt / M
19
Science programs ? momentum properties

• Schedules with long mean exposure times are worse
• More momentum buildup in a single observation
• Less opportunity to balance momentum
• Schedules with more constrained observations are
  worse
• Less ability to schedule at zero roll
• Less flexibility in overall schedule

Dump vs. fixed orient
40 ksec Isotropic zero roll if possible
20
Scheduling algorithms ? momentum properties
Preference for scheduling early
No bias in where to schedule
ltmogt1.21 n-m-s/day
ltmogt0.08 n-m-s/day
21
Strategies Long-range planning

• Spread targets over the instantaneous field of
  regard
• Mix short and long observations
• Maintain good balance of unconstrained and
  constrained visits
• Limit visit windows of unconstrained visits so
  they occur during times of low momentum
  accumulation rates.
• Limit visibilities by magnitude of momentum
  management .
• Identify blocks of visits to manage the momentum
  over the entire block.

22
StrategiesShort Term Scheduling

• Assign rolls (of unconstrained visits) to
  minimize overall momentum
• Order visits to reduce momentum buildup.
• Significant reductions in momentum buildup can be
  obtained by both techniques

23
Impact of visit failure

• Visit failures cause schedule to move forward and
  can change momentum profile
• To avoid unplanned dumps (or gaps) caused by
  failures
• Carry out acquisition failure analysis to
  determine margin
• Use OP windows to prevent buildup exceeding
  limits
• Limit maximum assigned (or by placement) off
  normal roll 2
• Update OP as needed weekly

24
Summary

• Webb operations are well understood
• Operations will be efficient
• L2 environment
• Event-driven operations
• Momentum management is going to be needed.
• Management reduces momentum buildup factor 2-4.