Constrained and Coordinated Observations

1
Constrained and Coordinated Observations

• Chandra Mission Planning Response to Questions
  from Chandra Users Committee

2
Statistics on Constraints
How many observations are actually constrained,
and in what category (e.g., approved by peer
review, preferences expressed in forms,
constraints added by users after their programs
were approved, DDT observations, etc.)?

• The statistics are summarized in the following
  table. Specific items to note are
• Nominally, the limit for constrained
  observations is supposed to be 20. This
• is intended to mean that constrained
  observations should not exceed 20 by
• number or time, but things are usually tallied
  by number at peer review.
• Most DDT observations are constrained in some
  way. Some are the equivalent
• of TOOs while others are the equivalent of
  somewhat right time windows and/or
• monitoring programs.
• Tracking constraints (or preferences) that have
  been added after peer review
• can be difficult, and some such constraints
  are satisfied without the constraints
• ever making it into the database.
  Operationally this probably seems larger than
• it really is.
• There is some double-counting of constrained
  observations in the database
• compilation because of splitting of ObsIDs.
  This should be relatively small.

3
Statistics on Constraints
How many observations are actually constrained,
and in what category (e.g., approved by peer
review, preferences expressed in forms,
constraints added by users after their programs
were approved, DDT observations, etc.)?
AO TOT TOTT PRC TPRC AC TAC PRP AP DDT DC DP TOTC
1 1025 22.2 188 18 4.1 18 107 3.6 20 12 7.6 34
2 852 21.5 236 28 6.6 31 69 1.4 34 22 8.0 37
3 929 22.5 225 24 6.0 27 76 1.5 4 46 30 5 7.6 34
4 889 20.0 169 19 4.3 21 5 0.1 63 5 23 18 5 4.4 22
5 830 23.9 279 37 5.5 23 37 1.1 42 2 14 11 5 6.9 29
AO Observing Cycle TOT Total number of
targets TOTT Total observing time (Ms) PRC
Targets w/ peer-reviewed constraints
(includes CAL and GTO) TPRC Time for targets w/
peer-reviewed constraints AC Number of
targets w/ added constraints TAC Time for
targets w/ added constraints
PRP Targets w/ peer-reviewed preferences
(includes CAL and GTO) AP Number of
targets w/ added preferences DDT Directors
Discretionary Time targets DC DDT
observations w/ constraints DP DDT
observations w/ preferences TOTC Total
constrained time (not including
preferences)
4
Constrained and Coordinated Observations
What is the appropriate overall number of
constrained observations (health and safety vs.
scientific value), and how many should be given
out by the peer review vs. DDT? This should
include a discussion of the load on CXC planning
and operating staff.

• Constrained observations do not, in general,
  represent a health and safety issue.
• - Constraints are identified in advance and
  evaluated at several points in planning
• process such observations are done safely or
  not at all.
• - TOO/DDT observations are more an issue for
  health and safety since these
• require rapid replanning and detailed
  review. These are a large manpower load.
• The primary costs for constrained observations
  are
• 1. Difficulty and loss of flexibility in
  scheduling
• - All constraints (and most preferences)
  are met in initial long-term schedule.
• - Competition between observing
  constraints, spacecraft constraints, visibility.

5

• Constraint Checks
• Roll violations
• Sunblock
• Phase constraints
• Windows
• Radzones
• Pitch angle
• Coordinations
• Monitoring

6
Constrained and Coordinated Observations
What is the appropriate overall number of
constrained observations (health and safety vs.
scientific value), and how many should be given
out by the peer review vs. DDT? This should
include a discussion of the load on CXC planning
and operating staff.

• Constrained observations do not, in general,
  represent a health and safety issue.
• - Constraints are identified in advance and
  evaluated at several points in planning
• process such observations are done safely or
  not at all.
• - TOO/DDT observations are more an issue for
  health and safety since these
• require rapid replanning and detailed
  review. These are a large manpower load.
• The primary costs for constrained observations
  are
• 1. Difficulty and loss of flexibility in
  scheduling
• - All constraints (and most preferences)
  are met in initial long-term schedule.
• - Competition between observing
  constraints, spacecraft constraints,
  visibility.
• - Future modifications difficult w/ lots of
  constraints (particularly if long duration).
• - Constraints added after LTS is built are
  a burden.
• - SOT/MP team has been reduced from 5 to 3
  scientists, with 5 data aides.
• ? Increasing number of constraints will
  become a manpower issue

7
Constrained and Coordinated Observations
What is the appropriate overall number of
constrained observations (health and safety vs.
scientific value), and how many should be given
out by the peer review vs. DDT? This should
include a discussion of the load on CXC planning
and operating staff.
2. Competing/Conflicting space for TOO/DDT
observations - LTS is heavily subscribed
with constrained observations. -
Interruption of schedule for TOO or DDT often
means working around or moving a
constrained observation. - Long
constrained observations are particularly
difficult in this regard, as are long
TOO/DDT observations.
8
Chandra Long-Term Schedule
9
Constrained and Coordinated Observations
What is the appropriate overall number of
constrained observations (health and safety vs.
scientific value), and how many should be given
out by the peer review vs. DDT? This should
include a discussion of the load on CXC planning
and operating staff.

• 2. Competing/Conflicting space for TOO/DDT
  observations
• - LTS is heavily subscribed with
  constrained observations.
• - Interruption of schedule for TOO or DDT
  often means working around
• or moving a constrained observation.
• - Long constrained observations are
  particularly difficult in this regard,
• as are long TOO/DDT observations.
• 3. Efficiency and conflicts w/ spacecraft
  constraints
• - As spacecraft constraints become more
  restrictive, constraints are more
• difficult to accommodate. Example
  EPHIN pitch constraints and need
• for cool targets. (This now dominates
  scheduling difficulties.)
• - Minimizing slew time is more difficult
  with large numbers of constrained
• observations. Pre-launch simulations
  indicated 20 was ok we aim for that.
• Current number of constraints awarded at peer
  review is probably a bit high.
• Higher levels would certainly be a burden.

10
Constrained and Coordinated Observations
What is the appropriate overall number of
constrained observations (health and safety vs.
scientific value), and how many should be given
out by the peer review vs. DDT? This should
include a discussion of the load on CXC planning
and operating staff.

• Time available for DDT observations 1 Ms for
  Cycle 1-4 700 ks in Cycle 5
• - full allotment not used in Cycles 1, 2, 4
• - large single segments (500 ks) used in
  Cycles 1, 2 not highly constrained
• - overall percentage of constrained budget is
  this not large primary difficulty is
• more along lines of TOO-like nature of many
  DDT observations

• Coordinated observations require more effort
  than typical constrained targets.
• - Planning/negotiation with planners from
  other observatories
• is manpower intensive (and often iterative).
• - Number per cycle is shown in table this
  does not include a fair
• number of preference coordinations that
  have been supported.
• - Pitch angle restrictions introduce
  additional complications
• because many observatories have pitch angle
  restrictions very
• near 90 degrees (Chandras bad pitch
  zone).
• - Increasing number beyond current levels
  would be a
• considerable burden.

AO
1 49
2 51
3 31
4 26
5 55
11
Changing Allowed Constraints and Formats
Can the constraints be changed for the ease of
users (e.g., allow setting time constraints in
GST)? The short answer is Yes, if necessary,
but it can be a large task. Heres why
RPS The proposal submission software requires
modification to accommodate new constraints or
new ways to specify existing constraints. Our
version of RPS is under CXC control and
modifications are performed by CXC staff.
Databases If modifications require new fields,
new ranges, or multiple values, the proposal
and/or observation databases require
restructuring. New elements must be added to
databases, software to populate/query these must
be modified, and many individual scripts need to
be updated.
Spike The long-term scheduling software has a
fixed set of constraints that it can handle.
New constraints require modification of the
software. Spike is maintained and modified by
STScI through a support contract.
OFLS The off-line system scheduling software is
used for generating the detailed schedule
(and also the spacecraft command loads). In
general, new constraints need to be added here as
well. Changes to this software are expensive, and
the turn-around time is very long.
To date, modifications have been made to support
multiple roll ranges, multiple window
constraints, grouping constraints, nonlinear
monitoring intervals, and preferences. All have
required significant software changes.
12
Changing Allowed Constraints and Formats

• Can the constraints be changed for the ease of
  users (e.g., allow setting
• time constraints in GST)?
• With specific regard to modifying the allowed
  constraints to support the
• specification of time constraints in Greenwich
  Sidereal Time
• Note first that Chandra does not support
  constraints for coordinated
• observations with ground-based observatories.
  In RPS, such a request can
• be made as a preference only.
• Addition of such support would require
  modifications to RPS, the Proposal
• database, ObsCat, Spike, and the OFLS.
  Modifications to the two scheduling
• software packages could be significant.
• While not technically supported, to date we have
  coordinated a significant
• number of observations with the VLA. In a very
  small number of cases,
• tight restrictions on the sidereal time have
  been required. These have been
• handled by hand. The number of such instances
  have been very small. A
• large software modification does not seem
  warranted for this.